Znorg. Chem. 1986, 25, 1461-1466 be ascribed to the stabilization experienced by an highly charged cation (Ni(II1)) in an environment of increasing negative charge (due to the increase of C104- ions). On increasing the NaC10, concentration, because of the opposite effects described above, the difference in stability of the Ni(II1) complexes with [9]aneN3 and cyclam, expressed by MI/,, which is quite small at low NaC104 concentrations (0.1 M Na= 25 mV), becomes larger and larger (7 M, 160 mV). C104, Moreover for NaC104 concentrations greater than 3 M (see Figure 6), the El~2(Ni(III)/Ni(II))value for the [ 16]aneNs complex becomes less positive than that of the cyclam analogue, thus inverting the previously established, apparently well-defined order.
Conclusions This work has tried to throw light into the factors that control
1461
the attainment of the trivalent state of nickel polyamine complexes in solution. At this stage, the following conclusions can be drawn. (1) The structural properties of the ligand (in particular the ring size) are a very selective element in the stabilization of Ni(II1); this is particularly evident in the case of quadridentate macrocycles. (2) Changing the solvent does not alter the relative trend of stability of Ni(II1) complexes with different macrocycles: the nature of the solvent is not selective. (3) In aqueous solutions the variation of the concentration of the so-called “inert” electrolyte introduces a novel element of selectivity, discriminating between redox changes involving or not involving water molecules. Acknowledgment. Financial support of the Italian Ministry of Education (MPI; 40%) is gratefully acknowledged.
Contribution from the School of Chemical Sciences, University of Illinois, Urbana, Illinois 61801, Corporate Research Science Laboratories, Exxon Research and Engineering Company, Annandale, New Jersey 08801, and Department of Chemistry, Stanford University, Stanford, California 94305
Reactions of MoS3, WS3, WSe3, and NbSe3 with Lithium. Metal Cluster Rearrangement Revealed by EXAFS Robert A. Scott,*la Allan J. Jacobson,lb Russ R. Chianelli,lb W.-H. Pan,lb Edward I. Stiefel,lb Keith 0. Hodgson,lc and Stephen P. Cramer*lb Received July 31, I985
EXAFS analysis has been used to study the structural changes that occur during the lithiation of the amorphous materials MoS3, WS,, and WSe, and the crystalline material NbSe3. For the three amorphous materials, an increase in the number of metal-metal bonds was observed, as well as a significant decrease in the metal-metal distance. A reduction in the number of metal-chalcogenide interactions was also apparent, along with an increase in the metal-chalcogenide distance. For Li,MoS,, the predicted Mo-Mo and Mo-S distances are 2.66 (3) and 2.50 (3) A, respectively. A reasonable model for this fully lithiated structure involves an octahedral Mo6 cluster analogous to those found in Chevrel-phasematerials. Similar clusters may be formed during lithiation of WS, and WSe,. Predicted W-W and W-S distances were 2.64 (3) and 2.47 (3) A for Li4WS,; for Li5WSe,,W-W and W-Se distances of 2.67 (3) and 2.61 (3) A were found. In contrast with that of the amorphous trichalcogenides, lithiation of crystalline NbSe3results in a slight contraction of the average Nb-Se bond length. From Se EXAFS, it was found that a significant Se-Se interaction persists in this material after lithiation, and it is difficult to say whether or not significant metal cluster formation occurs. These results have implications for interpretation of the electrochemical behavior of these materials.
Introduction The chemistry of molybdenum and tungsten combined with sulfur and selenium is extremely rich in its structural diversity and practical importance. The dichalcogenides (ME2: M = Mo, W; E = S, Se) are well understood structurally, and are important for their catalytic and electrochemical properties.2 The trichalcogenides (ME3) are amorphous materials, which have been shown to react readily with n-butyllithium and alkali-metal naphthalides to form amorphous compositions such as M’,MoS3 (M’ = Li, Na, K 0 Ix I4).3 Electrochemically, MoS3 cathodes in lithium cells show good reversibility provided they are not discharged further than the composition Li3MoS3.4 The amorphous structure of these material^,^ whether prepared by thermal or solution techniques,6 has hindered their characterization. The Chevrel phases M‘,Mo6E8 form a third clsss of molybdenum chalcogenides. These systems have been extensively studied because of their superconducting properties. Formally, the molybdenum oxidation state varies from +2 to +2*/, depending on the value of x , overlapping that observed in the most reduced of the amorphous trichalcogenide compositions. In a previous paper, EXAFS results on the untreated ME3 materials were described.’ Despite their lack of long-range order, local structure involving metal-metal bonding was observed, and a chainlike structure for MoS3 was proposed.* This paper describes the dramatic structural changes that occur upon electro*To whom correspondence should be addressed.
0020-1669/86/ 1325-1461 $01.50/0
chemical lithiation of these materials. Clear evidence is found for the formation of higher order metal clusters, and a hexanuclear octahedral model is proposed for Li4MoS3. It thus appears that the amorphous trichalcogenides are important intermediates that can be transformed to dichalcogenidesby heat or the Chevrel phase analogues electrochemically.
Experimental Section Sample Preparation and Data Collection. The ME, starting materials were prepared by previously described methods from the (NH4),ME4 precursors.’ Lithium was inserted into the trichalcogenidesat ambient temperature either by electrochemicalmeans or by reaction with n-butyllithium in dry hexane using published procedure^.^ Open-circuit voltages for 15 Li,MoS, samples prepared by reaction with n-butyllithium were measured for comparison with the previous electrochemical
(1) (a) University of Illinois. (b) Exxon Research and Engineering Co. (c) Stanford University. (2) Chianelli, R. R. Int. Reu. Phys. Chem. 1982, 2, 127-165. (3) Jacobson, A. J.; Chianelli, R. R.; Rich, S . M.; Whittingham, M . S . Mater. Res. Bull. 1919, 14, 1437-1448. (4) Jacobson, A. J. Solid Stare Ionics 1981, 5, 65-70. (5) Ratnasamy, P.; Rodrique, L.; Leonard, A. J. J . Phys. Chem. 1973, 77,
2242-2246.
(6) (a) Diemann, E.;Muller, A. Coord. Chem. Reu. 1973, I O , 79-122. (b) Chianelli, R. R.; Dines, M. B. Inorg. Chem. 1978, 17, 2758-2762. (7) Cramer, S . P.; Liang, K. S.; Jacobson, A. J.; Chang, C. H.; Chianelli, R. R. Inorg. Chem. 1984, 23, 1215-1221. (8) Liang, K. S.; de Neufville, J. P.; Jacobson, A. J.; Chianelli, R. R.; Betts, F. J . Non-Cryst. Solids 1980, 35-36, 1249-1254.
0 1986 American Chemical Society
1462 Inorganic Chemistry, Vol. 25, No. 9, 1986
2.5 0
g
-
Scott et al.
R
2.3
: L
2.1
5
...*
1.9
1.7 1.5
0
0.5 1.0
1.5
2.0
2.5
3.0
3.5
4.0
x in LI,MoS3 Figure 1. Resting potential of chemically lithiated MoS3 samples as a function of degree of lithiation (see text for details). 19980
Samples of Li,MoS3 were formed into cathodes by loading the solids into stainless-steel-mesh bags, which were subsequently coldpressed to ensure good electrical contacts. The cathodes were then immersed in a solution of LiCIO, (1.6 M ) in dioxolane and open-circuit voltages measured against a lithium foil cathode by using a high-impedance voltmeter. Cells were equilibrated for 1-3 days before voltages were recorded. All of the XAS spectra were recorded at the Stanford Synchrotron Radiation Laboratory in transmission mode with Si[220] or Si[ 11 I ] crystal monochromators. Samples were shipped to Stanford in sealed tubes under an argon atmosphere and loaded into sample cells in a nitrogen-filled Vacuum Atmospheres glovebox. The sample cells were kept in a He-purged Lucite box with Kapton windows during data collection. Data Manipulation and Analysis. The EXAFS was extracted from the absorption spectrum by previously described procedures,*” smoothed by Gaussian convolution, and Fourier filtered. A curve-fitting analysis of the filtered data utilized the expression x(k) =
Nb X-y
Aab(k) [exp(-2uab2k2)1 sin [2kRab
20020
Zoo00
20040
Energy (eV) Figure 2. Molybdenum K-absorption edge changes in MoS3 under various lithium loadings: untreated MoS, (-. -); Li2MoS, (- -); Li,MoS,
(-1. I
,
,
,
,
,
aab(k)l (1)
kRab
In this expression, Nb is the number of scatterers at distance Rabwith mean-square-distance deviation u,:. The total amplitude &(k) and phase shift a,,(k) have been described previously.’ The phase shift functions were obtained from the ME2 model compounds,’ by parametrizing the phase shift as aab(k) = a ,
+ a l k + a2km+ a,k”
(2)
and varying the parameters a, to obtain the best possible fit to the data. No improvement was observed when the phase shift was obtained by using a complex Fourier back-transformation of single shells of EXAFS data.I2 In all cases, the same threshold energy (Eo) was used throughout the analysis. The amplitude analysis for first coordination sphere interactions relied on tetrahedral ME?- models whenever possible. The amplitude function was parametrized as A(k) = co[exp(c,k + c2k2)]kC3
(3)
and the parameters ci were varied to obtain the best fit. When this expression could not yield a good fit, as in the case of backscattering by tungsten in the first coordination sphere, the complex Fourier backtransform was used to extract a total amplitude envelope. In either case, the overall amplitude was corrected by calculating b a b from the vibrational frequencies.10J1J3 Frequently, it is not possible to find a model compound in which the uab contribution is accurately known. In such cases, the results can be expressed as changes in ‘Tab between model and unknown. Alternatively, one can adjust the empirical amplitude to yield the same g a b value derived by using the theoretical amplitude functions and a scale factor. This approach is similar to the FABM (fine adjustment based on models) (9) Cramer, S . P.; Hodgson, K. 0.;Stiefel, E. I.; Newton, W. E. J . Am. Chem. Soc. 1978, 100, 2748-2761. (10) Cramer, S. P. In EXAFS for Inorganic Systems; Garner, C. D., Hasnain, S. S., Eds.; Daresbury Laboratory: Daresbury, U.K., 1981. (1 1) Cramer, S . P.; Wahl, R.; Rajagopalan, K. V.J . Am. Chem. SOC.1981, 103, 7721-1121. (12) Eisenberger, P. M.; Shulman, R. G.; Kincaid, B. M.; Brown, G. S.; Ogawa, S. Nature (London) 1978, 274, 30-34. (13) Cyvin, S . J. Molecular Vibrations and Mean Square Amplitudes; Elsevier: Amsterdam, 1968.
12640
12660
12680
12700
Energy (eV)
Figure 3. Selenium K-absorption edges: (a) untreated NbSe, (- -), Li,NbSe, (-); (b) untreated WSe, (- -), Li2WSe3(--), Li,WSe3 (-); (c) solid Fe2(Se2)(C0)6(- -), Li2[Fe2Se2(C0)6]in T H F solution (-). The Fe2Se2compounds were a kind gift from T. D. Weatherill and T. B. Rauchfuss and were prepared as described in ref 15. procedure of Teo et aI.I4 Both procedures have been used in the curve-fitting analysis of the lithiated trichalcogenides.
Results Oxidation State Changes upon Lithiation. The electrochemical insertion of lithium i n MoS3 under dynamic conditions has been reported p r e v i ~ u s l y .Results ~ u n d e r open-circuit conditions for the change in voltage as a function of composition for Li/Li,MoS, (0 I x I 4) electrochemical cells a r e shown in Figure 1. T h e initial sharp drop in voltage to 2.1 V is followed by a smooth falloff t o 1.86 V a t t h e composition Li3MoS3. Beyond this point t h e voltage drops more rapidly to t h e limiting composition Li,MoS,. Similar behavior is observed u n d e r d y n a m i c conditions, b u t t h e voltage profile is flatter and t h e falloff is m o r e rapid beyond 3 L i per MoS3. It has been observed t h a t cells discharged beyond t h e composition Li3MoS3 d o not recharge, indicating t h a t a n irreversible chemical c h a n g e occurs in this composition region. (14) Teo, B. K.; Antonio, M. R.; Averill, B. A. J . Am. Chem. SOC.1983, 105, 3751-3762.
Metal Cluster Rearrangement in Lithiated ME3
Inorganic Chemistry, Vol. 25, No. 9, 1986 1463
0
I
2
R'
0
1
3
2
4
5
(d)
R' Figure 4. Mo EXAFS Fourier transforms (k = 4-18 A-', k3 weighting) for MOSSwith progressively higher lithium loadings: (a) untreated MoS, (-), Li,MoS, (--); (b) Li2MoS3(-), Li3MoS3(--); (c) Li,MoS, (-), Cu, *Mo6SSChevrel phase (--).
The EXAFS results suggest that this irreversible change is to be associated with the formation of Mo-Mo-bonded metal clusters (vide infra). Treatment of M o S ~with lithium results in a progressive shift of the Mo edge to lower energies, as illustrated in Figure 2. This is consistent with the molybdenum becoming more reduced. Previously, it was proposed that, rather than Mov'(S2-)3, a better formulation for M o S ~would be M O ' ~ ( S ~ ~ - ) ( S ~ -From ) ~ . this formulation one would expect reduction of sulfur during lithiation. The fact that the molybdenum edge shifts more between LizMoS3 and Li4MoS3than between MoS3 and Li2MoS3may indicate that electrons initially reduce the disulfide bond preferentially. In order to better understand the electronic changes associated with lithiation, the Se edges of WSe3 and NbSe3 were examined as a function of lithium loading. As illustrated in Figure 3, profound changes in the height and shape of the Se edge were observed. Addition of two lithiums to WSe3 causes a significant decrease in the height of the main edge feature at 12658 eV, with a concomitant increase in intensity around 12 665 eV. Virtually identical changes are observed in the S e edge spectra of Fez(Se,)(CO), upon reduction to [Fe2Se2(C0),I2- as shown in Figure 3c. The latter reaction is known to involve the breaking of the Se-Se bond in a diselenide bridge.15 The Se edge spectra thus corroborate the suggestion that the initial reduction occurs at the dichalcogenide bond. Interpretation of the Se edge changes that occur upon conversion of Li2WSe3to Li5WSe3will have to await the study of other structurally characterized selenium compounds. Structural Changes upon Lithiation. The Mo EXAFS Fourier transforms for M o S ~with progressively higher lithium loadings are shown in Figure 4. Upon addition of a single lithium, the major change involves a decrease in the amplitude of the shorter distance Mo-S peak ( R ' z 1.9 A). The feature at R'zz 2.4 A, associated with a short Mc-Mo interaction, decreases in intensity somewhat between MoS3 and LiMoS3, but is lost almost completely between LiMoS3 and LizMoS3. These results are again consistent with the idea that the disulfide bond is reduced first. Further lithiation causes dramatic increases in the amplitude of the short-distance Mo-Mo feature in Li,MoS, and in Li4MoS3. It is clear that a significant structural rearrangement takes place during this part of the reaction of lithium with this material,
(A)
3
4
5
Figure 5. M EXAFS Fourier transforms ( k 3weighting, k = 4-18 A-'; except for WS, k = 4-16 A-1) for ME, materials before and after maximum lithiation: (a) untreated MoS, (--), Li4MoS3 (-); (b) untreated WS, (--), Li4WS3(-); (c) untreated WSe, (--), Li5WSe, (-); (d) untreated NbSe, (--), Li3NbSe3 (-).
R'
(A)
Figure 6. Se EXAFS Fourier transforms ( k 3 weighting, k = 4.0-16.0 for lithiated materials: (a) untreated WSe, (--), Li5WSe3(-); (b) untreated NbSe, (- -), Li3NbSe3(-). A-l)
resulting in fewer or more disordered Mo-S interactions and a greater number of short (R' z 2.4 A) Mo-Mo interactions. Similar structural rearrangements take place in WS3 and WSe,, as illustrated in Figure 5. In contrast, the crystalline material NbSe,, which is also electrochemically active, exhibited a smaller Nb-Se transform peak upon uptake of three lithiums. This may indicate a decreased or more disordered Nb-Se interaction, but it might also arise from destructive interference by a Nb-Nb interaction.I6 It was also possible to observe changes by using the Se EXAFS of WSe3 and NbSe3, and the Fourier transforms are shown in Figure 6. The tungsten triselenide data clearly indicate a longer Se-W interaction, whereas the NbSe3 EXAFS shows only a diminished Se-Nb interaction upon lithiation. In order to obtain a more quantitative interpretation of the structural changes associated with lithiation, curve-fitting analysis was required. In this analysis, three different variables had to (16) Other examples of coincidental destructive interferenceof EXAFS from
(1 5) Weatherill, T.D.;Rauchfuss, T.B.; Scott, R. A. Itwrg. Chem.,following paper in this issue.
different shells of scatterers have been documented. For example, see: Antonio, M. R.; Teo, B. K.; Averill, B. A. J . Am. Chem. SOC.1985, 107, 3583-3590.
1464 Inorganic Chemistry, Vol. 25, No. 9, 1986
Scott et al. Chart I
7\
M-M
M-M
I 4
I
7
I3
10
I 16
(2')
k
Figure 7. Curve-fitting analysis of fully lithiated ME3 materials using M EXAFS, showing filtered data (-) and fit (--): (a) Li,MoS3; (b) Li,WS3; (c) Li5WSe3,(d) Li3NbSe3.
4
12
8
16
k
Figure 8. Curve-fitting analysis of fully lithiated MSe3 materials using Se EXAFS showing filtered data (-) and fit (--): (a) Li,WSe3; (b)
Li3NbSe3. be adjusted: the coordination number &, the DebyeWaller factor exp(-2na:k2), and the interatomic distance Rab. The close correlation between Nb and n&,can cause large errors when both are adjusted simultaneously. On the other hand, dramatic changes in gab have been observed during cluster transformations," so use of a fixed nab cannot be justified. To test the stability of such a fitting procedure, the previously reported EXAFS spectra of several metal-sulfur clusters were reanalyzed according to the new procedures. The results for the ME3 starting materials and their lithiated analogues are summarized in Table I. Table I1 (supplementary material) contains the curve-fitting results on several structurally characterized model compounds. Although the distances were calculated with high accuracy, the metal-metal coordination numbers were sometimes in error by as much as +40% to -28%. These errors were not correlated with the type of amplitude function used; rather, they are intrinsic to the crude treatment of multiple-scattering effects. Hence, for structure predictions, it is necessary to use an error of *50% for the second-shell coordination numbers. First-shell errors are more likely f25%. Representative fits of the EXAFS of fully lithiated samples are illustrated in Figures 7 and 8. The numerical results for the lithiated samples are compared with revised analyses of (17) Cramer, S . P.; Eidem, P. K.; Dori, Z.; Gray, H. B. J . Am. Chem. Soc. 1983, 105, 799-802.
I
I1
I11
IV
the untreated ME, data in Table I. The MoS, curve-fitting results quantitate the drastic changes revealed in the Fourier transforms. The closest Mo-Mo interaction shortened by nearly 0.1 A, from 2.75 to 2.66 A. Meanwhile, the average Mo-S bond lengthened from 2.41 to 2.50 A. The coordination number changes are less precise, but the number of Mo-Mo interactions clearly increases while the average number of Mo-S bonds diminishes from -6 to 3 or 4. Similar results were obtained when the EXAFS spectra of lithiated WS3 and WSe, were examined. Fully lithiated WS3 showed a 0.1 1-A contraction in the W-W distance and an increase in the W-W amplitude. The W-S bond increased from 2.38 to 2.47 A. Upon full lithiation of WSe3, a 0.1-A reduction in the W-W distance and a 0.06-A lengthening of the W-Se bond were observed. However, the changes in coordination number were ambiguous. Contributing to this uncertainty is the fact that the W-Se and W-W distances have moved within 0.05 A of each other, so that the values obtained for the two shells are closely correlated. The S e EXAFS of fully lithiated WSe, confirmed this elongation of the Se-W distances and the decrease in the average number of Se-W interactions. It is difficult to deduce the structural changes occurring during NbSe3 lithiation, since no new transform features appear. However, it is clear that this crystalline material behaved differently from the amorphous materials. No dramatic lengthening of the Nb-Se distance was observed. If anything, there was a 0.02-A contraction to 2.66 A. Fitting the Se EXAFS yielded a Se-Nb distance of 2.62 A, as well as a Se-Se interaction at 2.39 A with a very large Debye-Waller factor. Discussion The EXAFS results on fully lithiated M o S ~ WS,, , and WSe, are consistent with an increase in the size of the metal clusters present in these materials. It is useful to consider some of the structures of species that may be involved and the bonding that may occur (Chart I). For the Mo-S system, examples of isolated discrete clusters are available for all four structural types. As the oxidation state of the molybdenum is lowered, the number of d electrons available for metal-metal bonding increases. Hence, the strength of the metal-metal interaction can increase via multiple bonding, or the size of the metal cluster can increase. An example of a simple dinuclear species as in I is the Mo(V) dimer [(Sz)2M~(S2)2Mo(Sz)2]2-, which has a Mo-Mo distance of 2.83 8, and a pair of disulfide bridges." A representative trinuclear species is the Mo(1V) trimer [ M O ~ ( S ) ( S ~ ) which ~]~-, has a triangular Mo, core with 2.72-A Mo-Mo bond lengths, a capping S2-ion, and three bridging SZz-ions.I9 Regular tetranuclear species are not common, but a recent example is the Mo(II1) cluster [Mo~S,(CN)~,]*-,which has a cubane-like MO4S4 core and an average Mo-Mo distance of 2.85 Finally, (18) Muller, A.; Nolte, W. 0.;Krebs, B. Angew. Chem., Inr. Ed. Engl. 1978, 17, 279. (19) Muller, A,; Pohl, S.;Dartmann, M.; Cohen, J. P.; Bennett, J. M.; Kirchner, R. M. Z. Naturforsch., B Anorg. Chem., Org. Chem. 1979, 346, 434-436.
Metal Cluster Rearrangement in Lithiated ME3
by a chalcogenide. One of the most symmetric Chevrel-phase compounds, Cu366MOgS8, has an average Mo-Mo distance of 2.67 similar to the Li,MoS3 Mo-Mo distance of 2.66 A. It is
Inorganic Chemistry, Vol. 25, No. 9, 1986 1465
1466
Inorg. Chem. 1986, 25, 1466-1472
clearity clusters most probably involves interchain interactions. A plausible transformation is illustrated in Figure 9. Here, a M3E4unit (half of M6Es) is built up from two metals in one chain and one metal in a neighboring chain. The chalcogenide that bridged the two chains in ME, becomes the “cap” of the M,E, unit. The other half of the M6Es cluster can be imagined to be contributed by another layer of interacting ME, chains. Of course, the EXAFS alone can shed no light on the nature of the intermediate structures involved in such a transformation. Perhaps other experiments could better address this point.
Summary EXAFS analysis has shown that lithiation of the amorphous materials MoS3, WS3, and WSe3 results in a condensation into larger metal cluster, which may be similar to the clusters found in Chevrel-phase structures. This result is consistent with chemical trends previously observed in molecular cluster chemistry. These lithiated materials represent a unique transition region between small molecular clusters and the solid state, and they may be of
synthetic value in the creation of large metal-chalcogenide clusters. Acknowledgment. The X-ray absorption measurements were performed at the Stanford Synchrotron Radiation Laboratory (SSRL), which is supported by the Department of Energy, Office of Basic Energy Sciences. The work was performed while R.A.S. was at Stanford as a N I H Postdoctoral Fellow. XAS work under K.O.H. is supported by the N S F and NIH. Registry No. M o S ~ ,12033-29-3; WS,, 12125-19-8; WSe,, 8898134-4; NbSe,, 12034-78-5; Li4MoS,, 76770-55-3; Li,WS3, 101011-07-8; Li,WSe,, 10101 1-08-9; Li,NbSe,, 55886-04-9; MoS,, 1317-33-5; [M02(S2)6]2-, 97278-58-5; [M03(S)(S2)6]2-, 88765-92-8; w s 2 , 1213809-9; WSe2, 12067-46-8; [WSe4I2-,21559-01-3; NbSe,, 12034-77-4; Se, 7782-49-2; [Fe2Se2(C0)6I2-,101011-09-0; Fe2Se2(C0),, 76185-26-7; Li, 7439-93-2; LiMoS,, 101011-10-3; Li2MoS3, 101011-1 1-4; Li3MoS3, 10101 1-12-5. Supplementary Material Available: Table 11, containing reanalysis of curve fitting for the structurally characterized compounds MoS,, [M02(%)6I2-, [ M O ~ ( S ) ( S Z ) ~w~s~2-, ,WSe2, [WSed2-, NbSe2, and Sea (3 pages). Ordering information is given on any current masthead page.
Contribution from the School of Chemical Sciences, University of Illinois, Urbana, Illinois 61801
Structural Evidence Concerning the Frontier Orbitals in [FezE2(CO),l2- (E = S, Se): Redox-Active Dichalcogen Ligands Timothy D. Weatherill, Thomas B. Rauchfuss,* and Robert A. Scott* Received October 18, 1985 Fe and Se extended X-ray absorption fine structure (EXAFS) data have been used to determine the structures of the Fe2E2(E = S, Se) cores of the following compounds: [Fe2E2(CO),]‘ ( z = 0, 2-), Fe,(p-ER),(CO), (R = alkyl), [Fe2E2(N0),]*-, and Fe2(p-ER),(NO),. For the carbonyl compounds, the Fe EXAFS results indicate that the Fe-C, Fe-Fe, and Fe-E distances change little upon conversion of Fe,(E,)(CO), to [Fe2E2(Co)6]2-and Fe,(p-ER),(CO),. The Se EXAFS results are consistent with the Fe EXAFS data, also showing that the two-electron reduction of Fez(Se2)(C0)6results in breaking of the Se-Se bond. The Fe and Se EXAFS of the related nitrosyls [Fe,E2(NO),l2- and Fe,(p-ER),(NO), confirm that the Fe2E2cores of these complexes are similar. All of the EXAFS results are consistent with the LUMO of Fe@2)(CO),, being mainly E-E antibonding in character.
Introduction Since its preparation in the mid-l950s,’ the compound Fez(S,)(CO), has been the subject of a number of investigations. Its solid-state structure as determined by Wei and Dahl features a tetrahedral Fe2Szcore, acute Fe-S-Fe angles, and a short Fe-Fe distance of 2.55 A.Z On the basis of this data, all atoms within the Fe,S, core are considered to be mutually bonded. Several theoretical studies have dealt with the bonding and electronic structure of this species, the nature of the Fe-Fe bond being the primary f o c u ~ . ~Wei , ~ and Dahl suggested that the structure of Fe2(S2)(CO), is due in part to the presence of a “bent” Fe-Fe bond. The origin of the iron orbitals responsible for this bond may be found by considering each iron atom as being octahedrally coordinated, the Fe-Fe bond arising from orbitals occupying the (sixth) position trans to the axial carbonyl ligands. Teo et al. supported this interpretation with a Fenske-Hall calculation and assigned the HOMO and LUMO to the bonding and antibonding components of the Fe-Fe i n t e r a ~ t i o n .An ~ ab initio calculation also describes the Fe-Fe interaction in terms of a ”bentn bond and assigns this molecular orbital as the HOMO.5 Self-consistent-field Xct scattered-wave molecular orbital calculations done by Andersen et al., provide a modified picture of the bonding
situation in Fez(s2)(CO)6.6 The SCF-Xa-SW results, together with a H e I photoelectron study, indicate that the S2 bridge enhances and stabilizes the r-component of the Fe-Fe bond. The assertion by Andersen et al. that the Fe-Fe bond possesses some multiple-bond character has been disputed by DeKock and coworkers on the basis of He I and He I1 photoelectron spectroscopy and Hartree-Fock%later calc~lations.~They argue that the LUMO in Fe2(S2)(CO),is antibonding with respect to the sulfur atoms, not the iron atoms. The chemistry of the dianion [Fe2S,(CO)6]z- suggests that the LUMO in Fe2(S2)(CO), is indeed primarily S-S antibonding8 Thus, Seyferth and co-workers have demonstrated that the reactivity of [ Fe2S,(Co)6]2- is sulfur-localized. Alkylation and metalation of this dianion afford neutral Fe2(pSR),(CO), derivatives: the integrity of the Fe-Fe bond is retained in the products (eq 1).
Hieber, W.; Gruber, J. 2.Anorg. Allg. Chem. 1958, 296, 91. Wei, C. H.; Dahl, L. F. Inorg. Chem. 1965, 4 , 493. Wei, C. H.; Dahl, L. F. Inorg. Chem. 1965, 4 , 1. For related theoretical discussions see: Burdett, J. K. J . Chem. Soc., Dalton Trans. 1977, 423. Summerville, R. H.; Hoffmann, R. J . Am. Chem. SOC.1976, 98, 7240. Mason, R.; Mingos, D. M. P. J . Organomer. Chem. 1973, 50, 5 3 . Teo, B. K.; Hall, M. B.; Fenske, R. G.; Dahl, L. F. J . Organomef. Chem. 1974, 70, 413. Teo, B. K.; Hall, M. B.; Fenske, R. F.; Dahl, L. F. Inorg. Chem. 1975, 14, 3103. Van Dam, H.; Louwen, J. N.; Oskam, A.; Doran, M.; Hillier, I. H. J . Electron Spectrosc. Relat. Phenom. 1980, 21, 57.
Although the structures of these alkylated and metalated products are firmly established, the structure of the parent dianion remains unknown. Herein we report the determination of the
0020-1669/86/1325-1466$01.50/0
( 6 ) Andersen, E. L.; Fehlner, T. P.; Foti, A. E.; Salahub, D. R. J. Am. Chem. SOC.1980, 102, 7422. (7) DeKcck, R. L.; Baerends, E. J; Hengelmolen, R. Organometallics 1984, 3, 289. DeKock, R. L.; Baerends, E. .I. Hengelmolen, ; R. Inorg. Chem. 1983, 22,4158. (8) Seyferth, D.; Henderson, R. S. Organomerallics 1982, 1 , 125.
0 1986 American Chemical Societv
